@bitkyc08/opencodex 2.7.31 → 2.7.33

package/src/types.ts

@@ -211,7 +211,7 @@ export type AdapterEvent =
   // the SAME output index, so the activity animates instead of flashing completed instantly.
   | { type: "web_search_call_begin"; id: string }
   | { type: "web_search_call_end"; id: string; queries: string[]; status?: "completed" | "failed"; sources?: OcxUrlCitation[] }
-  | { type: "done"; usage?: OcxUsage }
+  | { type: "done"; usage?: OcxUsage; stopReason?: string }
   // `usage` carries best-effort partial consumption when a turn dies before a clean done
   // (e.g. cursor upstream 502 mid-stream), so failed requests can log real token counts.
   | { type: "error"; message: string; usage?: OcxUsage };
@@ -347,6 +347,24 @@ export interface OcxClaudeCodeConfig {
   visionSidecar?: { backend?: "openai" | "anthropic"; model?: string };
 }
 
+/** 사용자가 대시보드에서 직접 추가한 커스텀 모델 정의. */
+export interface OcxCustomModel {
+  /** 고유 ID (crypto.randomUUID()) */
+  id: string;
+  /** 프로바이더 키 (기존 providers[name]) */
+  provider: string;
+  /** 모델 슬러그 (프로바이더 접두사 없는 bare id) */
+  modelId: string;
+  /** 인간 가독 표시명 (선택, 슬래시 불가) */
+  displayName?: string;
+  /** 컨텍스트 윈도우 (토큰) */
+  contextWindow?: number;
+  /** 입력 모달리티 (선택, 기본 ["text"]) */
+  inputModalities?: string[];
+  /** 추가 시각 (ISO 8601) */
+  addedAt?: string;
+}
+
 export interface OcxConfig {
   port: number;
   providers: Record<string, OcxProviderConfig>;
@@ -402,6 +420,8 @@ export interface OcxConfig {
    * are omitted from the bare /v1/models list.
    */
   disabledModels?: string[];
+  /** 사용자가 대시보드에서 직접 추가한 커스텀 모델 목록. */
+  customModels?: OcxCustomModel[];
   /**
    * Shadow call intercept: redirect Codex Desktop's hard-coded gpt-5.4-mini helper calls
    * (title generation, commit messages, skill orchestration) to a user-chosen model.
@@ -585,6 +605,15 @@ export interface OcxWebSearchSidecarConfig {
   routedModelStallTimeoutMs?: number;
 }
 
+export interface OpenRouterProviderRouting {
+  /** OpenRouter provider slugs to try first, in priority order. */
+  order?: string[];
+  /** Restrict routing to these OpenRouter provider slugs. */
+  only?: string[];
+  /** Whether OpenRouter may use providers outside `order`. Defaults to OpenRouter's policy. */
+  allowFallbacks?: boolean;
+}
+
 export interface OcxProviderConfig {
   adapter: string;
   baseUrl: string;
@@ -633,6 +662,10 @@ export interface OcxProviderConfig {
   /** Model-specific max input token limits. Values cap auto_compact_token_limit. */
   modelMaxInputTokens?: Record<string, number>;
   headers?: Record<string, string>;
+  /** Default provider-routing preferences for models sent through the canonical OpenRouter API. */
+  openRouterRouting?: OpenRouterProviderRouting;
+  /** Exact model-id overrides for `openRouterRouting`. Each matching entry replaces the default. */
+  modelOpenRouterRouting?: Record<string, OpenRouterProviderRouting>;
   /**
    * "key" (default): authenticate upstream with `apiKey`.
    * "forward": relay the caller's incoming auth headers verbatim (OAuth passthrough; gpt only).
@@ -692,6 +725,12 @@ export interface OcxProviderConfig {
    * only on explicit `true`. See devlog/_plan/260709_parallel_tool_calls.
    */
   parallelToolCalls?: boolean;
+  /**
+   * Opt-in: forward `prompt_cache_key` to the upstream `/chat/completions` body.
+   * OpenAI-specific extension; strict backends (Groq, Cerebras, etc.) reject unknown
+   * fields. Default off; only enable for providers that document this parameter.
+   */
+  promptCacheKey?: boolean;
   /** Model ids whose tool_choice only accepts `auto` or `none`; forced/named choices are downgraded. */
   autoToolChoiceOnlyModels?: string[];
   /** Model ids that expect prior assistant `reasoning_content` to be preserved in chat history. */
@@ -768,6 +807,8 @@ export const OPENAI_PROVIDER_TIER_VERSION = 2 as const;
 export interface CodexAccount {
   id: string;
   email: string;
+  /** User-owned display label; never participates in routing or identity checks. */
+  alias?: string;
   plan?: string;
   chatgptAccountId?: string;
   logLabel?: string;

package/src/update/index.ts

@@ -1,4 +1,4 @@
-import { spawnSync } from "node:child_process";
+import { spawn, spawnSync } from "node:child_process";
 import { readFileSync, readdirSync } from "node:fs";
 import { fileURLToPath } from "node:url";
 import { dirname, join } from "node:path";
@@ -260,7 +260,24 @@ export async function runUpdate(): Promise<void> {
           windowsHide: true,
         });
         if (svcStdio === "pipe") logSpawnOutput("", svc);
-        if (svc.status !== 0) console.warn("⚠️  Service refresh failed — run 'ocx service install' manually.");
+        if (svc.status !== 0) {
+          // On Windows, schtasks /create requires elevation. The CLI inherits the
+          // user's (non-admin) token, so the service reinstall can fail with access
+          // denied. Fall back to a direct detached proxy start so the update never
+          // leaves the user without a running proxy.
+          console.warn("⚠️  Service refresh failed — starting the proxy directly instead.");
+          console.warn("   Run 'ocx service install' as administrator to refresh the background service.");
+          const env = { ...process.env };
+          delete env.OCX_SERVICE;
+          const child = spawn(process.execPath, [process.argv[1], "start", "--port", String(capturedListen.port)], {
+            detached: true,
+            stdio: "ignore",
+            windowsHide: true,
+            env,
+          });
+          child.unref();
+          console.log(`✅ Proxy starting on port ${capturedListen.port}.`);
+        }
       } finally {
         if (prevBake === undefined) delete process.env.OCX_BAKE_PORT;
         else process.env.OCX_BAKE_PORT = prevBake;

package/src/update/job.ts

@@ -330,17 +330,27 @@ async function restartAfterUpdate(
     }
     const prevBake = process.env.OCX_BAKE_PORT;
     process.env.OCX_BAKE_PORT = String(Math.trunc(port));
+    let serviceOk = false;
     try {
       const run = io.runService ?? ((j, bin, args) => runLoggedCommand(j, bin, args, RESTART_TIMEOUT_MS));
       const result = run(job, cmd.bin, cmd.args);
-      if (result.status !== 0) {
-        throw new Error(`service restart failed (${cmd.display}, exit ${result.status ?? "?"})`);
+      serviceOk = result.status === 0;
+      if (!serviceOk) {
+        // On Windows, `schtasks /create` requires an elevated token. The update worker
+        // inherits the (non-admin) proxy's privileges, so a service-managed install
+        // updated from the GUI or a normal terminal fails here with access denied.
+        // Falling back to a direct proxy start keeps the update from leaving the proxy
+        // stopped; the stale service manager can be refreshed later with an admin
+        // `ocx service install`.
+        updateJob(job, {}, `Service reinstall failed (exit ${result.status ?? "?"}); falling back to a direct proxy start. Run 'ocx service install' as administrator to refresh the background service manager.`);
       }
     } finally {
       if (prevBake === undefined) delete process.env.OCX_BAKE_PORT;
       else process.env.OCX_BAKE_PORT = prevBake;
     }
-    return;
+    if (serviceOk) return;
+    // Fall through to the direct proxy start below so the update never leaves the
+    // proxy stopped when the service reinstall could not run.
   }
 
   const pid = readPid();

package/src/usage/expected-prices.ts

@@ -37,6 +37,7 @@ const KIMI_K27_CODE: Cost4 = { input: 0.95, output: 4, cacheRead: 0.19, cacheWri
 const KIMI_K27_CODE_HIGHSPEED: Cost4 = { input: 1.9, output: 8, cacheRead: 0.38, cacheWrite: 1.9 };
 const KIMI_K26: Cost4 = { input: 0.95, output: 4, cacheRead: 0.16, cacheWrite: 0.95 };
 const KIMI_K25: Cost4 = { input: 0.6, output: 3, cacheRead: 0.1, cacheWrite: 0.6 };
+const QWEN38_ROUTEWAY_TEMPORARY: Cost4 = { input: 1.5, output: 5, cacheRead: 0.15, cacheWrite: 0 };
 
 const GEMINI_PRICING = "https://ai.google.dev/gemini-api/docs/pricing (2026-07-22); cacheWrite=0: storage is billed per-hour, not per-token";
 const MINIMAX_PRICING = "https://platform.minimax.io/docs/guides/pricing-paygo";
@@ -44,6 +45,9 @@ const DEEPSEEK_PRICING = "https://api-docs.deepseek.com/quick_start/pricing-deta
 // Kimi official tables publish input/output/cache-hit only; cacheWrite is mapped to the
 // cache-miss input price (Kimi auto-caches with no separate write billing). 2026-07-20 re-verified.
 const KIMI_PRICING = "https://platform.kimi.ai/docs/pricing (official table; cacheWrite derived = input, Kimi auto-cache has no write billing)";
+// TEMPORARY proxy only: Routeway's reseller API rate is not Alibaba Token Plan billing.
+// Replace these overlays when Alibaba publishes an official qwen3.8-max-preview token rate.
+const QWEN38_ROUTEWAY_PRICING = "https://routeway.ai/models/qwen3.8-max-preview (temporary reseller proxy; NOT Alibaba Token Plan billing; cacheWrite unpublished -> 0)";
 
 export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   // MiniMax M2.1 highspeed — published PAYG price (verified).
@@ -55,6 +59,8 @@ export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   // Google Antigravity effort-suffix variants — derived from the verified base-model
   // price (Google does not publish per-suffix prices; Agent inference bills at the
   // base model's standard rate per the official Billing FAQ).
+  { provider: "google-antigravity", modelId: "gemini-3.6-flash", cost4: GEMINI_36_FLASH, source: `collapsed base ID ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified" },
+  { provider: "google-antigravity", modelId: "gemini-3.1-pro", cost4: GEMINI_31_PRO, source: `collapsed base ID ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified" },
   { provider: "google-antigravity", modelId: "gemini-3.1-pro-low", cost4: GEMINI_31_PRO, source: `derived: gemini-3.1-pro (<=200k tier) ${GEMINI_PRICING}`, verifiedAt: "2026-07-20", status: "verified-derived" },
   { provider: "google-antigravity", modelId: "gemini-3.1-pro-high", cost4: GEMINI_31_PRO, source: `derived: gemini-3.1-pro (<=200k tier) ${GEMINI_PRICING}`, verifiedAt: "2026-07-20", status: "verified-derived" },
   { provider: "google-antigravity", modelId: "gemini-3.6-flash-low", cost4: GEMINI_36_FLASH, source: `derived: gemini-3.6-flash ${GEMINI_PRICING}`, verifiedAt: "2026-07-22", status: "verified-derived" },
@@ -95,6 +101,10 @@ export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   { provider: "kimi-code", modelId: "kimi-k2.6", cost4: KIMI_K26, source: KIMI_PRICING, verifiedAt: "2026-07-20", status: "verified-derived" },
   { provider: "kimi-code", modelId: "kimi-k2.5", cost4: KIMI_K25, source: KIMI_PRICING, verifiedAt: "2026-07-20", status: "verified-derived" },
   { provider: "kimi-code", modelId: "kimi-for-coding", cost4: KIMI_K27_CODE, source: `derived: kimi-k2.7-code ${KIMI_PRICING}`, verifiedAt: "2026-07-20", status: "verified-derived" },
+  // Alibaba has not published a per-token Token Plan rate yet. Use Routeway's
+  // independently published reseller rate temporarily and keep estimates derived.
+  { provider: "alibaba-token-plan", modelId: "qwen3.8-max-preview", cost4: QWEN38_ROUTEWAY_TEMPORARY, source: QWEN38_ROUTEWAY_PRICING, verifiedAt: "2026-07-22", status: "verified-derived" },
+  { provider: "alibaba-token-plan-intl", modelId: "qwen3.8-max-preview", cost4: QWEN38_ROUTEWAY_TEMPORARY, source: QWEN38_ROUTEWAY_PRICING, verifiedAt: "2026-07-22", status: "verified-derived" },
   // Cursor Auto router — Cursor's published fixed token price (verified).
   { provider: "cursor", modelId: "auto", cost4: { input: 1.25, output: 6, cacheRead: 0.25, cacheWrite: 1.25 }, source: "https://docs.cursor.com/account/pricing + https://cursor.com/blog/aug-2025-pricing", verifiedAt: "2026-07-20", status: "verified" },
 ];

package/src/usage/summary.ts

@@ -63,6 +63,7 @@ export interface UsageModel {
   inputTokens: number;
   outputTokens: number;
   shareRatio: number;
+  estimatedCostUsd?: number;
 }
 
 export interface UsageProvider {
@@ -74,6 +75,7 @@ export interface UsageProvider {
   estimatedRequests: number;
   totalTokens: number;
   shareRatio: number;
+  estimatedCostUsd?: number;
 }
 
 export interface UsageSummary {
@@ -350,6 +352,29 @@ function buildModels(entries: PersistedUsageEntry[], totalTokens: number): Usage
       else if (status === "estimated") model.estimatedRequests += 1;
     }
   }
+  // Accumulate per-model estimated cost
+  for (const entry of entries) {
+    const estimate = entry.attempts?.length
+      ? estimateComboCost(entry.attempts)
+      : estimateRequestCost({ provider: entry.provider, model: entry.model, usage: entry.usage, usageStatus: entry.usageStatus });
+    if (!estimate) continue;
+
+    if (entry.attempts?.length && estimate.attempts) {
+      // Combo: attribute each attempt's cost to its own model
+      for (const attemptEst of estimate.attempts) {
+        const aProviderKey = baseProviderLabel(attemptEst.provider);
+        const aKey = `${aProviderKey}${attemptEst.model}`;
+        const m = byKey.get(aKey);
+        if (m) m.estimatedCostUsd = (m.estimatedCostUsd ?? 0) + attemptEst.cost.total;
+      }
+    } else {
+      // Single-target: attribute to the entry's model
+      const providerKey = baseProviderLabel(entry.provider);
+      const key = `${providerKey}${entry.model}`;
+      const m = byKey.get(key);
+      if (m) m.estimatedCostUsd = (m.estimatedCostUsd ?? 0) + estimate.cost.total;
+    }
+  }
   const models = [...byKey.values()];
   for (const m of models) m.shareRatio = totalTokens === 0 ? 0 : m.totalTokens / totalTokens;
   return models.sort((a, b) => b.requests - a.requests);
@@ -396,6 +421,24 @@ function buildProviders(entries: PersistedUsageEntry[], totalTokens: number): Us
       else if (status === "estimated") provider.estimatedRequests += 1;
     }
   }
+  for (const entry of entries) {
+    const estimate = entry.attempts?.length
+      ? estimateComboCost(entry.attempts)
+      : estimateRequestCost({ provider: entry.provider, model: entry.model, usage: entry.usage, usageStatus: entry.usageStatus });
+    if (!estimate) continue;
+
+    if (entry.attempts?.length && estimate.attempts) {
+      for (const attemptEst of estimate.attempts) {
+        const aProviderKey = baseProviderLabel(attemptEst.provider);
+        const p = byKey.get(aProviderKey);
+        if (p) p.estimatedCostUsd = (p.estimatedCostUsd ?? 0) + attemptEst.cost.total;
+      }
+    } else {
+      const providerKey = baseProviderLabel(entry.provider);
+      const p = byKey.get(providerKey);
+      if (p) p.estimatedCostUsd = (p.estimatedCostUsd ?? 0) + estimate.cost.total;
+    }
+  }
   const providers = [...byKey.values()];
   for (const p of providers) p.shareRatio = totalTokens === 0 ? 0 : p.totalTokens / totalTokens;
   return providers.sort((a, b) => b.requests - a.requests);